MicroSave...As the chart shows, ASA began opening urban branches in 2001, and the number has grown steadily. One additional function that urban branches perform is to members who migrate

MicroSave Market-led solutions for financial services

Offices across Asia, Africa and Latin America

www.MicroSave.net [email protected]

Managing Growth of MFIs: ASA Bangladesh -

single-minded growth

Stuart Rutherford1

with Graham A.N. Wright and S.K. Sinha

- Final- February 2008

1 During the preparation of this study, ASA provided office space, and access to its staff and documents. This made our work easy and we thank them. ASA clients and other villagers and slum dwellers gave us their unstinting time and attention, and to them we are especially grateful. The lead author thanks David Cracknell and Graham Wright at MicroSave for inviting him to contribute this paper to their series on ‘Managing Growth’. The views expressed here are those of the authors and are not necessarily shared by ASA or by MicroSave.

http://www.microsave.net/�

mailto:[email protected]�

INTRODUCTION

SA has not been the fastest growing of the MFIs in this collection of studies. Al Amana of Morocco and Compartamos of Mexico, for example, have had faster year-on-year growth rates both in terms

of clients served and loans disbursed and outstanding. But ASA has been at it longer – since 1992 – and is now remarkable for its sheer size. At the end of 2007 it is serving about 6.7 million clients, of whom about 5 million are borrowing, their total loans outstanding amounting to about US$350 million (at market exchange rates: $1.4 billion at purchasing-power parity). Among Asia’s specialist pro-poor microlenders, only Grameen Bank and BRAC – both Bangladeshi and both older than ASA – are larger. The purpose of this series of studies is to review the difficulties faced by rapidly-growing MFIs and to examine how they were addressed. It is hoped that this will provide younger MFIs with a rich understanding of the sort of problems they are likely to face and the range of solutions they might try. But curiously, and despite its sixteen years of accumulated experience, the ASA story may not have the most to contribute. This is because the range of problems that ASA has had to face as a microlender has been unusually narrow. It did not have to fight for and establish the idea and the methods of microcredit among the public, donors, regulators, and potential clients. That task had already been done, principally by Grameen and BRAC, both of whom had been developing microcredit since 1976 – sixteen years before ASA turned to it. By 1992 most villagers had become accustomed to MFIs in their midst. The NGO Affairs Bureau, the government office which oversaw NGO microlenders like BRAC and ASA, was busy encouraging NGOs to run microcredit schemes and encouraging donors to provide funds for the purpose. Nor did ASA have to deal with institutional birth-pangs while it was developing microcredit. It had been founded in 1978, and worked first in rural ‘consciousness raising’ (forming poor villagers into groups and helping them to find ways to fight the injustices that beset them) and then in the delivery of rural services in health, law and education, and in disaster relief and rehabilitation. By the time ASA turned to microcredit it already had a large staff deployed in many districts. It had a legal identity that allowed it to do microcredit, and a secure leadership which had successfully overcome rivals and has never since been challenged. It had many contacts in the bureaucracy. It had already enjoyed many years funding from a variety of donors, private and bilateral. Bangladesh is a young country with a turbulent history. Its politics have been marked by violence and disorder, as it vacillates between democracy and martial law. But for microcredit providers the business environment has been surprisingly stable. Succeeding governments have not interfered much. The rest of the business sector, including formal banking, has left it alone. ASA has neither had to change its legal identity, nor chosen to. It has remained an NGO throughout, so has not faced the challenges of transforming to a for-profit entity. Its funding has evolved smoothly: first came generous donors, then some small commercial bank loans, then low-cost loans from a subsidised wholesaler, and finally, as outreach expanded and surpluses piled up, client deposits and retained earnings. But by far the most important of the stable conditions in which ASA has prospered has been the expanding demand from an enormous client pool for the very basic service that ASA offers. Though it was originally imagined (and is still marketed internationally) as microenterprise credit, standard Bangladeshi microcredit – as offered by nearly all the many hundreds of MFIs in the country that follow Grameen’s lead, as ASA does – is a general-purpose credit facility. Loans are offered in modest values: the first loan is about the equivalent of one month’s household income, and for most borrowers subsequent loans do not rise to more than two or three times that ratio. The loans are repaid in equal weekly instalments, and because the loan term is usually one year the instalments are small enough to be affordable from normal household cash-flow, making the loan easy to repay irrespective of how it is used. Very conveniently, from the point of view of the customers, the lender comes to the village each week to collect the instalments. Experience has shown that the very poor (those whose cash flow is so weak that they cannot find a modest fixed sum each and every week) and the well-to-do (those who find the weekly routine too cumbersome for such modest loans) do not participate in the schemes – or, having participated and found

A

Managing Rapid Growth of MFIs: ASA Bangladesh – Stuart Rutherford

MicroSave – Market-led solutions for financial services

ii

it uncomfortable or tedious, soon drop out. That still leaves a truly massive number of poor-to-middle-income households – probably between 15 and 20 million – who, in the absence of other good ways to manage their money, are likely to find the service useful all or much of the time. Whether they invest in a business, buy or lease land, repair or furnish their homes, educate or marry their children or bury their parents, deal with sickness and emergencies, indulge in treats, or just buy food and fuel and clothing, there is always something worth taking these easy-to-pay loans for. ASA’s growth has taken place over the period during which this basic service was being rolled out by an increasing number of MFIs to most of the villages and slums of Bangladesh. Though ASA has experimented with and expanded its products, it did so as a matter of choice, not of necessity: the core service has remained the low-value year-long weekly-repayment loan, the staple of its successful growth. ASA has not had to undergo large-scale internal reorganisation or training because the basic product and its delivery have remained largely unchanged. None of this is to disparage ASA’s achievement. Blessed with advantages, ASA found ways to exploit them to the full. Aside from Grameen and BRAC, no other organisation has ridden the tide of Bangladeshi microcredit as successfully as ASA. Forbes magazine was not far wrong when, late in 2007, it placed ASA at the top of its list of the world's 50 best MFIs.

*** 0 * 0 * 0 *** We will describe some of the ingredients of that success. The paper starts with a statistical review of ASA’s growth and ends with some commentary on what the future may hold, but is otherwise laid out according to the standard format for this series of papers provided by MicroSave. In addition, we have been asked to review an early and influential paper on ASA’s growth by Pankaj Jain: that review appears as an Appendix.

Contents

INTRODUCTION ............................................................................................................................................................. i

ASA’S GROWTH ........................................................................................................................................................... 1

MANAGING GROWTH AT ASA: ISSUES AND STRATEGIES .......................................................................................... 6 1 Organisational form and governance ................................................................................................................. 6 2 Growth strategy .................................................................................................................................................... 8 3 Product strategy ................................................................................................................................................. 11 4 Product delivery ................................................................................................................................................. 14 5 Operational management .................................................................................................................................. 15

PAST, PRESENT AND FUTURE: AN OVERVIEW .......................................................................................................... 19

APPENDIX: REVISITING THE PANKAJ JAIN PAPER ................................................................................................... 22



1

ASA’S GROWTH

Clients

n common with almost all Bangladeshi MFIs, ASA refers to its clients as members. This reflects not only their organisation into groups, but also the continuing preference, stemming from microfinance’s

origins in social development NGOs, to think of the relationship between microfinance institutions and their client as more than a merely commercial one. Chart 1: growth of membership and of active borrowers, 1992-2007 Thousands of members; left scale; % female borrowers, right scale

-

1,000

2,000

3,000

4,000

5,000

6,000

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 20070.0%

20.0%

40.0%

60.0%

80.0%

100.0%

borrowers non-borrowers % female borrowers

The surge in the number of non-borrowers in 2005 (and the dip in the proportion of borrowers who are female) occurred because in that year ASA started recruiting the husbands and fathers of its women clients as ‘guardian members’. They are eligible for small loans that act as supplementary loans to those held by their wives and daughters. They do not need to attend the weekly meeting, and they repay their loans monthly. By no means all of them choose to borrow, but the proportion who do has steadily increased, to half of them in 2007. In common with all Bangladeshi MFIs, the turnover of clients is extremely high. In an average month in mid 2007, for example, ASA was recruiting about 150,000 new clients but losing about 130,000, for a net gain of 280,000 in the year. Many of the ‘new’ clients are not new to microfinance: many have been clients of MFIs in the past, or still are – having an account in several MFIs is very common. Indeed, many of the ‘new’ clients held accounts one or more times in ASA itself, but dropped out.

I



2

Outlets ASA’s field offices were first called Units, later Branches. With 1,500 to 1,600 group members served by a Manager and up to five Loan Officers, an ASA branch serves a somewhat smaller area than the Bangladesh MFI average, allowing staff more immediate access to its members. Chart 2: growth of branches, 1992-2007

Offices

Head Quarters District* Region* Branch

* District and Regional offices share premises with branches

As the chart shows, ASA began opening urban branches in 2001, and the number has grown steadily. One additional function that urban branches perform is to serve members who migrate from the rural to the urban areas: ASA has an elaborate system under which rural branch staff track such members and inform the urban branches. Staff More of ASA’s staff are field-based than those of its main rivals. Moreover, a bigger proportion of the field-based staff have daily contact with clients. Although ASA has developed a traditional multi-layered organisation in the field – with Regional and District level posts – these senior posts do not have separate offices, but share space with branches. Chart 3: growth of staff, 1992-2007

-

5,000

10,000

15,000

20,000

25,000

30,000

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

male female

Charts 1, 2 and 3 all clearly show the spurt in growth that occurred after 2003.

-

500

1,000

1,500

2,000

2,500

3,000

3,500

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

rural urban



3

Loan portfolio Chart 4: growth of loan portfolio, 1992-2006 millions taka

-

5,000

10,000

15,000

20,000

25,000

30,000

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007-

10,000

20,000

30,000

40,000

50,000

60,000

outstanding at year end, left hand scale annual disbursements, right hand scale Most of the loan portfolio is in the standard core product – the small annual loan offered to women group members. It accounted for 74% of the entire loan portfolio at the close of 2007. Deposits ASA began with small uniform compulsory weekly savings, and these have grown steadily in value. It has twice experimented with commitment savings devices. More recently, it has introduced compulsory ‘loan security’ insurance. Chart 5: growth of client deposits, 1992-2006 millions taka: includes savings, insurance, and member security funds

0100020003000400050006000700080009000

10000

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 20070%

10%

20%

30%

40%

50%

60%

70%

80%

annual deposits, left hand scaleyear-end balance, savings + insurance + security, left hand scaleannual growth in balance, right hand scale

The chart vividly shows the transaction intensity of these savings: the annual growth in the balance of savings held by ASA is quite low, but each year very large sums are deposited and withdrawn. Many of the withdrawals result from the large numbers of account closures. Withdrawals by continuing clients are used for all manner of purposes, including loan repayment.



4

Source of funds Donations from bilateral and private international sources had constituted ASA’s funds in the years before it turned to microfinance in 1992. This resource funded its entry into microfinance. Thereafter donations continued, were supplemented by bank loans and then, increasingly, by client deposits and retained earnings. From the late 1990s ASA also received substantial subsidised funding from the national microfinance fund wholesaler, PKSF, but has been paying down those loans steadily since 2003. Chart 6: growth of funds, 1992-2007 Year end loan fund sources, millions taka

0

5,000,000

10,000,000

15,000,000

20,000,000

25,000,000

30,000,000

35,000,000

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

bank loans donors PKSF deposits retained earnings other

Performance ASA has become the most profitable MFI in Bangladesh. For the year 2005 – the latest year for which we have comparative data at ‘The Microbanking Bulletin’ – ASA’s return on assets was 14.53% (compared to the average for large Asian MFIs of 4%) and its operational self-sufficiency2

-

500

1,000

1,500

2,000

2,500

3,000

3,500

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007-

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

surplus, m taka, left hand scale yield, %, right hand scale

254.88% (compared to 131% for large Asian MFIs). Chart 7: growth of returns, 1992-2007 Surplus (millions of taka) and Effective yield (interest income as % of average portfolio)

2 For a definition of these terms see The MIX Market, at www.mixmarket.org. The ratios quoted come from MIX. The data for 1992-2006 in Chart 7 come from ASA’s audited financial statements. Note that these are not adjusted by the international standards which are intended to make comparisons between MFIs possible by pricing grants as if they were commercial loans. They show the surpluses that ASA earned each year end as a result of deploying its resources, including accumulated grants.

http://www.mixmarket.org/�



5

ASA made an (unadjusted) surplus from every year since 1992, though for the first five years, before the effects of scale boosted income, they were modest. The dip in 1998 results partly from the severe floods that Bangladesh suffered that year, and the drop in 2007 over 2006 largely from the ‘Sidr’ cyclone of November 2007. ASA has always expressed its loan interest as a ‘flat’ rate. The rate was 15% (approximately 32% APR) until July 1995 when it dropped to 12.5% at a time when MFIs were coming under increasing criticism in the press for their prices. The rate was raised again to 15% in January 2001, following floods and cyclones, and stayed there until the beginning of 2007 when they fell back to 12.5% again as ASA ‘gave some of the benefit’ back to its members. These rates are average for Bangladesh: somewhat higher than those of the Grameen Bank but on a par with the rate (of 12.5%) recommended to its partners by PKSF.



6

MANAGING GROWTH AT ASA: ISSUES AND STRATEGIES3

1 Organisational form and governance ASA’s legal form and governance have not been significant factors in its growth. Its founder’s sharp focus on a single goal is clearly reflected in its economical and decentralised organisational structure. Legal form: ASA is registered as a non-profit social-welfare organisation under the Societies Act, and is permitted by the government’s ‘NGO Bureau’ to receive foreign donations. Under new legislation that came into effect in early 2007 ASA has also gained registration with the Microfinance Regulatory Authority (MRA), a new entity housed in the central bank. Bangladesh’s microfinance legislation is weak, and is looking increasingly old-fashioned. Unlike other countries in Asia (Cambodia and Pakistan are good examples) it offers no ‘enabling’ legal form to microfinance providers who want to mobilise deposits from the general public4

Governance: The driving force behind ASA has always been its energetic founder, Shafiqual Haque Choudhury, who has been its CEO throughout, now with the title of President. He is the arbiter of virtually all decisions within ASA, whether they originate in his own mind or filter up from staff and clients. The current Chairman of the seven-person Governing Body, of which Choudhury is the ex-officio Secretary and agenda-setter, recognises that her role and that of her colleagues is to advise, monitor and approve, rather than initiate, policy

or to work as a for-profit. The currently available forms condemn MFIs to the arguably weak governance structures typical of NGOs. It is not yet clear how the new MRA will use the powers given it by the legislators. Recently, they have drafted a set of proposed rules which include an interest rate cap of 12.5% (flat); the standardization of savings and loan products; requiring MRA permission to open branches or to mortgage any property and (possibly) outlawing contractual savings products. The draft has been strongly opposed by MFIs and the outcome remains uncertain. Historically, legal form has not been an issue for ASA. Its registration as a Society suited its original social-development aims, and by the time it began work in microcredit in 1992 it had become clear that such registrations confer the option to do so. Indeed, the government’s own NGO Bureau was, at that time, encouraging NGOs into microfinance. Because the Bureau oversaw applications for use of foreign donations this meant in effect that the government was also encouraging overseas donors to provide funds for microcredit. Most Bangladeshi MFIs, including ASA, focus on microcredit, so the lack of licence to mobilise savings has not much concerned them. This is particularly so because by an informal agreement the authorities are happy for MFIs to take compulsory (and some voluntary) deposits from their group members, taking the view that most such members are net borrowers most of the time. Bangladesh’s government has not intervened in financial services for the poor as strongly as, for example, neighbouring India which has long favoured interest-rate caps and has required banks to serve poor households with prescribed products. NGOs and their donor-backers have therefore enjoyed wide scope to pursue their interests in Bangladesh. It is doubtful whether the microfinance movement would have grown so strong so quickly if this had not been so. Nevertheless when ASA went beyond the informal agreement on savings and offered deposit services to people outside its borrowing groups it was obliged to backtrack. At the time this was seen as a minor setback (see the later sub-section on product strategy). But things may change. As competition between MFIs sharpens, Grameen’s advantage in being legally able to raise deposits is beginning to matter more and more, and ASA may start to press for a legal form that allows it to do the same. We explore this a little more in the last section.

5

3 This, the longest section of the paper, is laid out according to the framework provided by MicroSave 4 Grameen Bank’s ‘bank ordinance’ is an exception: it was a one-off drafted uniquely for Grameen. 5 We thank Mrs Taherunnessa Abdullah for a private interview in Dhaka in November 2007

. Under the Societies Act the Governing Body members serve as unpaid volunteers and have extremely limited liability for the actions of the executive. They meet



7

regularly, are well respected, and are seen by Choudhury as helpful providers of advice – he frequently consults them privately between meetings. Nevertheless, at least in the last decade or so, they have never sought to materially change any of Choudhury’s policies. The one challenge to Choudhury’s predominance, mounted by a disgruntled staff member, occurred in the 1980s6

There is no training. No training cell, no training centre, no trainers. Work routines are standardised and simplified so that new recruits need only a few days of supervised work experience in a branch before being sent off to another one to start work. Head office staff are given no in-service training.

, before ASA went into microfinance. The then Chairman of the Governing Body played a decisive mediating role, coming down in Choudhury’s favour. The incident led to a revision – away from stakeholder democracy towards a more conventional and stable system – in the way the Governing Body and the CEO are selected. Organisational form: ASA has a head office based in the capital-city, Dhaka, with a staff of 220 (a little less than 1% of the total payroll of 24,700). It is headed by a President (Choudhury) and four Executive Vice Presidents. They determine policy and products, manage resources, and set the rules for and monitor the field. The field is a hierarchy composed of districts, regions and branches – of which there are now 3,300. This sounds conventional, but there are some surprises. For example:

Districts and regions have no support staff and no separate offices of their own. District and regional managers are supervisory staff who share a building and services with one or more branches.

Branches have no accountants. Accounting and cash-handling is simplified, distributed between the branch manager and the three or four loan officers, and then subjected to a tight schedule of repeated monitoring by senior staff at four different levels stretching up to head office. Nor do they have guards: male staff live on the branch premises.

Branch managers distribute loans but have little say in the decision to award the loan. This is not because that function is reserved for more senior staff but for the opposite reason: the work manual sets out the conditions under which loans are made and in what value, so there is hardly a decision to be made. The combination of delegation, standardisation and repeated monitoring shrinks the opportunities for rent-seeking by staff.

ASA has no divisions for vehicle maintenance nor property. It owns only four cars, field officers buy and maintain their own motorbikes7

The keywords are simplicity, standardisation, and decentralisation. ASA has been shaped by Choudhury to execute a single function in the most cost-effective way to achieve a single clear goal. The function is to disburse and collect microcredit, and the cost-effectiveness comes from refusing to spend money on anything that does not contribute directly to this function. The goal is to maximise client numbers, and the clarity comes from refusing to see microcredit as anything other than getting loans out to as many poor people as possible and then recovering them. The approach is given symbolic form in the ‘ASA self-reliant microcredit model’ – a cash-flow chart that maps exactly how the function leads to the goal. It focuses the entire organisation’s culture on rapidly expanding the loan portfolio through the relentlessly efficient use of resources, leading to ‘sustainability’ in the form of growing surpluses for reinvestment. ASA staff, in their day-to-day work, see this in even starker terms. For them, two benchmarks count more than any other: portfolio size and overdues. These are the measures that dominate the series of inspection visits that each branch receives and which comprise ASA’s monitoring ‘cascade’:

, and head office staff – even the EVPs – use public transport to reach the field. All premises other than the head office are rented.

twice monthly by the Regional Manager once quarterly by the District Manager (usually with the Regional Manager in tow) once in 2 to 3 years by a Director from head office (usually with the District and Regional

Managers in tow)

6 The story is told in Rutherford, ASA, the Biography of an NGO, ASA, Dhaka 1995, pages 72-73 7 They buy them at a subsidised price on an instalment plan.



8

These visits, which combine the filling up of set-piece formats with situation-specific discussions, never fail to bear down on the basics: getting loans out and – harder – getting them back. Ask any field-based staff what the hardest part of their work is and you will be told ‘loan collection’. They are likely to tell you that the next hardest is recruiting and keeping clients. Managing those two difficult tasks reasonably well is the key to keeping the seniors happy. The ‘fit’ between corporate goals and worker behaviour could hardly be closer: ASA is a machine crafted so that everyone’s interests are served by disbursing as many loans as possible and getting them all recovered. Its architecture follows from that.

2 Growth strategy When ASA moved to microcredit it did so swiftly, totally and ruthlessly, compressing the need to handle culture change to a brief moment. There was no piloting. ASA’s new business plans consisted of the practical arrangements made to carry out the most simple of visions: maximising the growth potential of branch profitability as expressed in the ‘Self Reliant Model’. Planning is short-span and budgeting adjusts constantly to the process of matching plans proposed from the field to resource availability. Business planning: In the late 1980s ASA – above all Choudhury – became dissatisfied with and anxious about the course it was following. It was accepting funds from overseas donors and deploying them in whatever rural services suited the donor’s preferences. In some places education was the main service on offer. In others it was health, or legal aid. When a natural disaster occurred – floods and cyclones mostly – there were opportunities to carry out relief and rehabilitation work. ASA never had fewer than seven international donors in the period 1985-1990, and the flow of grants rose each year: 10 million taka in 1987, 22 million in 1988, 31 million in 1989 and 37 million – at that time worth about US$1.48 million – in 1990. However, Choudhury chafed at this dependency on donors (and the reliance on natural disasters to open the donor purses). In an extraordinary paragraph in an ASA Annual Report he wrote that

‘…dependency on external resources reduces efficiency and quality. As… foreign funds… become easy for the NGOs so people [that is, NGO staff] get wages out of less effort.’

Moreover, he recognised that ASA’s work was having little or no impact. He wrote that,

‘after several years of involvement with the poor… ASA as well as the group members realised their development is delayed….’

But by that time Choudhury had an alternative plan in mind. He had watched as ASA members deserted his groups in droves, joining Grameen and BRAC to access credit. ‘People’, he wrote ‘realise that there is

The branch blackboard Perhaps nothing illustrates ASA’s direct and simple ‘common-sense’ approach as the branch blackboard, found in every branch since 1992 (and duplicated on the screens of the computers now being installed in branches). Each morning the blackboard displays the name of each officer alongside the repayments and savings they are expected to bring back after visiting their groups. To make this work, products have to be sufficiently standardised so that cash flows can be accurately predicted. The effect is to focus staff attention on the basics – getting loans repaid and savings collected.



9

no alternative to credit intervention for increasing their income and thus reducing poverty.’ In 1987 and 1988 ASA had tried giving loans to flood victims, as a way of making donor money go further, and had found that even in chaotic post-disaster situations poor people showed a strong propensity to repay loans. After years of decrying credit, they began to look more carefully at it. Choudhury was impressed, but felt that the credit providers had not yet got it right: they had discovered how to lend systematically to the poor, but failed to follow up the implications that this held for the operations and long-term viability of their institutions. The ‘business plan’ that this led to has scarcely changed since. At its heart is the ‘ASA Self-Reliant Model’, a simple bar graph showing how a single ASA branch can break even and start delivering surpluses within a year of its foundation. The version shown here dates from 2000, but apart from some tinkering with the values to take care of inflation, it is identical both to the version shown in the 1995 book ASA: the Biography of an NGO (page 101) and the most recent Annual Report (2006, page 27). ASA uses it frequently in newspaper articles and other publicity. To the extent that ASA has a formal business plan it consists of the accumulated decisions made by its senior staff to turn this graph into reality, reinvesting the surpluses to multiply the number of branches, and turning ASA into an entirely self-reliant entity independent not only of donors but even of commercial creditors. The model leads directly to standardisation. Only if all the numbers in the graph are under control – all branch expenses, the flow of loans, repayments and savings, the flow of capital between the branch and head office – and only if the growth rate of a branch is similarly under control – dates for the recruitment of members and the staged deployment of staff – can the outcome be certain. Once the surplus has been generated, there has to be a foolproof mechanism for employing it, hence the rigorous rules requiring branches to on-lend cash directly to other branches once a given cash limit is reached. Happily, the environment of rural Bangladesh is favourable to this kind of standardisation. Though there are regional differences in wealth and economy they are narrow relative to those in other countries. Language and other cultural factors are homogenous enough to treat every branch area similarly. Despite this, of course, real life refuses to obey human dictates – even ASA’s. For that reason ASA’s planning horizon, at six months, is short8

Culture change: ASA has twice changed its goals. In the mid 1980s it shifted from ‘consciousness raising’ to service delivery, and then in 1991-2 to microcredit. In each case the change followed a similar

, so that plans can be quickly changed to suit circumstances. For the same reason, budgeting is a mix of top-down and bottom-up. Routinely, field-staff develop 6-month forecasts of what they hope to achieve, but when these plans reach the head office and are consolidated and matched with resources, they may be approved, or districts may be instructed to look for additional opportunities to expand, or told to cut back. ASA’s growth strategy is ‘extensive’ in that it has depended heavily on adding identical branches. But to the extent that improvements are constantly sought to the way that the ‘self-reliant model’ is operationalised, it has ‘intensive’ elements. Recently, for example, staff recruitment has been largely transferred from head office to the districts, enabling ASA to attract a wider group of potential applicants at less cost.

8 Though longer term scenarios are sometimes built to help senior staff think about future options.



10

pattern. After a period of anxiety and rising realisation that goals were not being achieved, Choudhury made up his mind and then acted swiftly. For the shift into microcredit this meant, among other things, a double sex change: the men in ASA’s rural beneficiary groups were bluntly told to go home and send their wives or daughters the following week. ASA had observed that BRAC and Grameen did much better with women, who were seen as more tractable and ‘obedient’, than with men, who were restless in the weekly meetings if they bothered to turn up at all. ASA thus provided itself overnight with a readymade and docile microcredit client base. At the same time ASA sacked most of its locally-recruited women workers and employed new non-local men, believing that outsiders to a locality would be less likely to favour certain households and – men especially – less open to pressure from local elites. The new microcredit programme was brought in everywhere, at once: ASA does not practise piloting. Head office simply brazened out any staff complaints, set staff to work with a busy new schedule, and waited for things to settle down, which they did. In the case of the change to microcredit some older staff regretted that the more egalitarian and utopian aspects of ASA had been finally laid to rest, but they were outnumbered by new staff who had never known anything except the new dispensation. There were no serious problems. For the sixteen years that have passed since then there has been little need to manage further changes to the institutional culture. ASA did bring in new products, including new savings products which are often thought to require sensitive culture change management, but these innovations were small relative to the massive core loan product. Moreover, as we shall see in a succeeding section, they were often less ‘flexible’ than they at first seemed, and could be slotted easily into the familiar daily routines at the branches. Where they were not thus, they could be quickly dropped with little resulting damage. It is possible that changing ground conditions may soon require ASA to face the need to make another fundamental change to its goals and culture – we explore that in the final section. But setting that aside, the most obvious lesson that ASA offers to organisations that need to make fundamental changes is ‘think about it carefully, then when you’ve made up your mind carry it through quickly and ruthlessly’. Corporate identity and resource acquisition: Choudhury started ASA while he was working for CCDB, a Christian relief and development NGO that was set up in the wake of Bangladesh’s 1971 war of independence. His work there put him in touch with western donors – mostly churches or NGOs – and he was able to use the presentation skills he learned in CCDB to get some of those donors to fund a new and more revolutionary kind of NGO: ASA. When ASA turned away from revolution to service delivery, it continued to use Choudhury’s skills to attract overseas donor funding: by the end of the 1980s, as we have seen in an earlier section, ASA had seven donors and annual grant income exceeding US$1 million. When ASA became a microcredit NGO, these skills remained important at the beginning (see chart 6). In 1992 and 1993, just after the shift to microcredit, ASA received its biggest ever grant income. But Choudhury wanted out of donor-dependency and into self-reliance. ‘Self-reliance’ began to define ASA’s public image. Conveniently, it had two applications, covering ASA’s developmental and institutional goals – self-reliance for the rural poor, achieved through borrowing productively from ASA; and self-reliance for ASA, marking ASA out as a new kind of leaner, harder-working, more efficient NGO-MFI. In 1992 ASA became one of the first NGOs to borrow from PKSF, a government-owned wholesaler of microcredit funds that used both government money and large grants from the World Bank. Indeed for sometime ASA was the biggest single taker of PKSF funds. A little later, in 1995, ASA became the first Bangladesh NGO-MFI to take a loan at commercial rates (9% a year) from a bank, when it took $250,000 from the Agrani Bank, secured by its Dhaka head office. This deal – hard to bring off and helped along by USAID’s Financial Sector Reform programme – further established ASA as an unusually ‘businesslike’ NGO. In other ways ASA has handled its public image cautiously and shrewdly. It has been better-than-average at getting articles into local newspapers, and been consistent in ensuring that these always trumpet its core virtue of self reliance for both the members and for ASA. As we shall see in a later section, customer service has not been a key selling point of Bangladeshi microcredit, and ASA does not use its branch offices to present a distinctive image: like most other NGOs, its branch buildings are simple rented structures. But it does line the roads with more signboards than most others, and it does it



11

consistently, using a standard format, colouring and logo. In 2006 it completed and moved into its own 14-storey ‘ASA Tower’ building in Dhaka, but in this respect it was not a trendsetter – it was the last of four big MFIs (Grameen, BRAC, and Proshika are the other three) to build high. Like most NGOs, it has stayed aloof from Bangladesh’s fiery politics. One large NGO-MFI shrivelled after a disastrous entry into party politics, whereas ASA avoided association with any political parties, and Choudhury’s one brief appearance in a political role, in early 2007, was as an Advisor to the neutral Caretaker Government. 3 Product strategy ASA is a microcredit provider, and its core product has not changed very much over 16 years of business: a high-volume low-value basic general-purpose loan, repaid in weekly instalments. Though ASA has experimented with other loan types, none looks set to replace the small loan. The experiments have not much disrupted ASA’s business, and the same can be said of its experiments with savings. Savings are used largely to protect the small loan product, by providing some de facto security against default, rather than as a way of raising capital. ASA’s case shows that an MFI can grow large without needing branch-level computerisation, providing its products and their delivery is straightforward enough, and that shifting to IT in middle-age, as ASA is doing right now, can be rather uncomfortable. Standardised loans: The workhorse of ASA’s growth has always been the small annual loan, repaid weekly: at end 2007 the small loan accounted for 75% of the loan portfolio. ASA has depended much more on a growing its standard product than on product diversification aimed at attracting a wider group of clients and doing a wider range of business with them. This made sense in the Bangladesh context, where the public argument for microcredit was precisely that it did not seek a wide range of clients but focused on the poor, and where BRAC and Grameen, among others, had already shown that there was huge demand for a basic loan service that could be used for the many and varied needs facing poor people with low-value household cash flows. ASA’s strategy of entering this market with a more efficient delivery system enabled it eventually to claim as big a share of microcredit clients as BRAC and Grameen. The market proved so large that its leaders are only now being forced to think more seriously about what to do as it matures. From the later 1990s onwards ASA experimented with other loan products. Most fall into one of three classes: higher-value loans for small businesses, special products for the very poor (similar to the core product but with lower values and more relaxed repayment schedules), and supplementary loans for the small loan users. None of these trial products has proved as durable as the small annual loan, and none has enjoyed such good portfolio quality. Take, for example, the history of the business loan. Two versions were introduced in 1997: the SEDP (small entrepreneurship development program) loans for men, and ‘individual credit’ for both men and women. A year later the ‘individual credit’ approach was quietly dropped. In 1999 the SEDP was renamed the ‘Small Trading’ loan and its target client group redefined, and then changed again, to ‘Small Business’ loan the following year. By the end of that year, 2000, there were 65,000 small business borrowers with 528 million taka outstanding between them (about 8,000 taka or about $116 each), compared to the small loan programme which had 3.45 billion taka outstanding among 1.1 million borrowers (an average of about 3,000 taka or $43 each). By 2003 the number of small business borrowers had barely grown, and ASA introduced another version – the SEL (small entrepreneur lending program), targeted at somewhat bigger businesses. Midway through 2007 these two products – Small Business and SEL – held, between them, a 15.4% share of the total portfolio. Their payments in arrears have tended to be a little higher than for the small loan. For future prospects, see the final section. Savings, a secondary service: It has been argued9

9 See for example, Robinson, Marguerite S., "Introducing Savings Mobilization in Microfinance Programs: When and How ?", Microfinance Network, Cavite, Philippines, and HIID, USA (November 1995).

that introducing voluntary savings into microcredit programmes throws up many challenges, and may require re-engineering the entire culture of the MFI. That has not yet been much of an issue in Bangladesh, where microcredit has always used savings – compulsory savings made only by members of borrowing groups – linked in some way to the loans, but few have tried independent voluntary savings accounts. Although ‘income generating’ uses of loans were being promoted as the best way that borrowers could be sure of repaying their debt, compulsory savings



12

were being collected and used – in one way or another – to part-guarantee the loans. These deposits soon accumulated to become a substantial contributor to funds for lending. The NGOs collected the weekly deposits alongside the loan repayments, and often supplemented the saved total by compulsory deductions from loans at disbursement. The debate in Bangladesh, as savings has crept up the agenda in microfinance discussions, has been about the best mechanisms for drawing down these savings and, more recently, the extent to which voluntary savings can supplement or replace the compulsory savings. That debate can be seen very clearly in ASA’s case. ASA began with fixed-value weekly deposits – fixed for all the members of a 20-person group. These savings could not be withdrawn until the member left the group. This is a typical standardised ASA solution: every week the expected total deposit is known in advance (and is chalked up on a branch blackboard at the beginning of the working day), and the loan officer is charged with the simple task of ensuring it is collected. The efficiencies of this system are clear: it simplifies accounting and forecasting, contributes stable funds that can be lent back to the members, builds up a cushion against default, and cuts down on the scope for irregular behaviour by loan officers. On the other hand its disadvantages soon became equally clear: not everyone wants to save at the same rate, and the more the deposits build up, the louder is the demand from savers to be allowed to withdraw some or all of them10

However, for the last few years members may make withdrawals of up to 500

. ASA has been tinkering with solutions to these two problems. It has trialled ways of allowing members to save at rates that differ from each other or vary over time: the current solution is to allow each member to save at any rate she chooses but to agree to save at that rate each week. Withdrawals policy has also been experimented with and the current policy is to allow withdrawals of any amount but such that a fixed proportion of the last-disbursed loan remains on account. These solutions still allow ASA to predict the weekly deposit total at the meeting.

11 taka (about $7) at the weekly meeting. This clearly breaches the ‘predictability rule’: the practice of designing products such that the flow of cash from the field each day can be predicted (and made public on the branch backboard). This worries some ASA staff. As a result, some loan officers try to dissuade members from making on-site withdrawals, or may request them to give advance notice. The resulting ad hoc experiment in running a flexible passbook savings regime is as close as ASA gets at present to abandoning the predictability rule, and it may have done so in the expectation that, with increasing competition, product rules in Bangladesh are set to become more flexible, and ASA may have to learn how to manage them. There is a precedent: in 1997 ASA introduced ‘associate savings’. This was a facility targeted at the relatives of group members and was fully open-access. It was closed soon afterwards, ostensibly because the Central Bank made it clear that while it would turn a blind eye to non-bank NGOs collecting savings from borrower-group members, collecting from non-group members by would be considered a breach of banking law. While some in ASA maintain that associate savings would have continued had the central bank not signalled its disapproval in this way, others take a view closer to the one set out in a MicroSave note12

Long-term savings are a different case. In these, the predictability rule is satisfied, because such products require fixed periodic deposits, and the withdrawal occurs only on the maturity of a fixed term, or if the account is abandoned. Long-term ‘contractual’ savings more obviously build capital for the institution, since short-term open access savings tend to feature very high transaction values but low balances. Long-term contractual savings accounts have always been very popular with middle-class Bangladeshis, and when in the 1990s BURO, a mid-size NGO-MFI, began offering them to its group members, evidence began to mount that they were popular with some, if not all, of the microcredit client group. This evidence turned into a torrent after Grameen introduced a similar product in 2002

: that the benefits to the institution of running open-access savings were not clearly greater than the difficulties it imposed on ASA’s field practices and book-keeping systems, and that it would have been closed down anyway.

13

10 In the case of Grameen, savers in central Bangladesh in 1995 arranged ‘strikes’ and succeeded in closing down many Grameen groups. Grameen responded by allowing savers to take back their savings when their membership reached ten years. Since then Grameen has become much more active in savings and is the MFI leader in savings in Bangladesh: see the MicroSave publication, Rutherford et al, Grameen II: The First Five Years. 11 1,000 taka in urban branches. 12 MicroSave Briefing Note 2: Introducing Savings into a MicroCredit Institution – Lessons from ASA 13 See Rutherford, Grameen II, op cit

. ASA has had two



13

tries at introducing it, the first in 1997 as part of that year’s flurry of experiments with savings, and then again more recently. At the close of 2007 it has a modest share of ASA’s total savings portfolio. However, there was a stronger reason than pleasing clients and building capital behind ASA’s renewed interest in contractual savings. It goes back to what was said at the beginning of this discussion: that savings have always been used as security against loan risk. Indeed, almost all ASA’s involvement in savings, no matter how it is presented, can be seen as ways of using savings to make credit work better. An early argument for allowing clients to withdraw savings was that clients might at some stage simply stop repaying credit if too much of their money was locked up for too long. The current policy of allowing members to withdraw savings at weekly meetings has certainly helped loan recovery, as clients often use the withdrawals to make loan repayments. Contractual savings further help to give ASA a cushion of savings to set against loans, and recent product rule changes make it clear that borrowers of more than a given loan value are expected to open a contractual savings account (as they are at Grameen, where client deposits now exceed client loan outstanding balances by three to two). The same calculus lies behind the system of ‘security funds’ that ASA runs. Under these schemes (there are different versions for members and for their husbands) members are required to deposit a fixed sum of 10 taka a week (this comes to about $7 a year), and another 10 taka for their husband. In return, ASA pays out six times the accumulated deposits when the member dies, and three times when the husband dies. These payouts are capped by limiting the term of deposits to 8 years: if the member is still in the system after 8 years she gets her fund back with annual interest of 4% (2% in the case of husbands). The fraction of outstanding loan that the average client holds in various savings and security funds is steadily rising, though, at 37% (end 2006), it is still much lower in ASA’s case than Grameen’s (150%). We return to this theme in the final section. IT systems: As we write this at the end of 2007 computers are being set up and tested in all of ASA’s 3,300 branches. Until now, ASA’s use of IT has been minimal. Branches sent monthly hand-written formats via their district managers to head office, where a team of data entry clerks typed them into the spreadsheets that comprised ASA’s consolidated management data. The shift to branch computerisation did not come about because ASA was outgrowing its manual system, which fitted both the philosophy and practise of its keep-things-simple approach rather well. And although computerisation will help ASA manage any new products that it may need to develop in the future in response to a putative demand for more flexibility from an increasingly sophisticated clientele, that does not seem to have been uppermost in senior management’s minds. Rather, head office wanted to release more branch staff time to tackling client matters (above all increasing arrears), and saw computerisation principally as a way of speeding up the record-keeping task without changing it in any fundamental way. For ASA, the manual system was iconic: it was Choudhury’s own invention and the perfect example of his talent for simplification. Its elegance enabled ASA to convince sceptical outsiders that running branches without accountants was both feasible and safe. By letting go of it, Choudhury has, for the first time, symbolically let go of his control over detail. Nevertheless, ASA is dealing with the IT decision as it does with all big decisions: having thought about it, it is going ahead, without piloting, in a massive programme over a short time span. A large team of developers was hired – 30 software writers housed in the head office backed up by 150 technicians in the field. In late 2007 ASA promoted 3,300 loan officers to a new post – assistant branch manager – responsible for maintaining the computers and training the other loan officers to use it, since the ambition is that each loan officer will make his or her own data entries, just as they now do with the manual system. All this is of course causing some disruption: probably more than any other decision in recent years. There are inevitable eddies of resentment and suspicion at head office (the software engineers are highly paid, and some ASA old hands feel that they have not been properly consulted about the shape of the programme). The promotion of 3,300 loan officers not only has significant cost consequences but entails the rapid recruitment of 3,300 fresh ones to take their place, triggering a shift in recruitment practices (see above). The knock-on effect has been that ASA has suspended its programme of branch expansion for six months – the first time this has ever happened.



14

Does ASA’s shift to IT have lessons for others? Probably not. That is not so much because the final outcome of the ASA experience is not yet known, but more because today’s younger MFIs will almost certainly start with a more sophisticated understanding of, and much better access to, IT, and will have dealt with in their initial plans. 4 Product delivery ASA establishes branches much as it does everything else – in a matter-of-fact and highly decentralised way. For the core product – the small annual loan taken by women – customer service has yet to become a competitive factor. Branch establishment and management: ASA’s way of establishing new branches typifies its general approach to its tasks. There is no team of analysts in head office poring over demographic and economic data and spying on the branch expansion plans of its competitors, no clever algorithms to help senior managers decide where to place branches. Not a bit of it. Head office does nothing except monitor the resources available. Everything else is done in the field. Every regional and district manager identifies locations where new branches may be feasible according to the simplest of criteria – the availability of a potential client pool of about 2,500 households within a few minutes bicycling distance. Their plans go to head office according to the six-monthly planning horizon discussed in an earlier section. If head office knows that resources are available, it lets the local managers go ahead with their plans undisturbed – they do not have to get head office permission for each branch (it helps, of course, that ASA is an NGO and not a bank – it needs no permission from a regulator to open branches). Only if resources are constrained or unexpectedly large for some reason will head office tell local managers to hold back or accelerate their plans. The simplest of formulae inform head office on available resources: key is the flow of excess liquidity from existing profitable branches. Similarly simple rules of thumb tell the district manager and the new branch manager how to proceed, such as ‘start with a skeleton staff and recruit so many clients and disburse so many loans before taking on further loan officers’. The route map is the ‘ASA Self-Reliant System’ that we illustrated on page 13, and its details are laid out for the branch manager in the Manual, a fat book of instructions which head office updates and sends out to each branch every two years. Branches are simple rented buildings, furnished identically according to rules set out in the Manual (only one fan, costing x, shared by all staff who sit around a table, dimensions a x b x c, centred under the fan – and so on). If a regional manager is based at the branch, there’s an extra chair; for a district manager an additional room, table, chair and filing cabinet. Male staff members live on the premises and act as guards. Women, who may work in their home areas, live at home or board locally. This system has worked without serious problems for sixteen years. There was only one instance of a disgruntled branch manager trying to capture his branch, and that was quickly and severely dealt with. Branches now number 3,300 and are found in every one of Bangladesh’s 64 Districts. Customer service: In Bangladesh microfinance customer service is not a fiercely contested competitive factor as it is in some other countries – and as it should be. MFIs do not advertise themselves on TV as a listening, or a caring, bank. Ideas about client-provider relationships have developed very slowly: clients are still conventionally seen – even by themselves – as grateful recipients of MFI beneficence rather than as keenly sought-after buyers of services. In the early days Choudhury once remarked to this writer that ‘if we wanted to, we could tell our members that they have to wear a red sari and stand on their heads in order to get a loan – and they’d do it’. That may have been an exaggeration even then, but competition has yet to dent its underlying truth in the way that it has impacted on other aspects of service – interest rates, length of wait before the first loan, and value of the first loan, for example. To this day clients call male NGO staff ‘sir’, give the first greeting when he arrives at the weekly meeting, and dutifully coil themselves at his feet and endure his barked reproofs (‘Selina, why are you late?’) Clients are not given printed product rule sheets – they rely entirely on what the loan officer chooses to tell them for their knowledge about the products they are using. The social conventions being thus, loan officers have little difficulty establishing themselves as the dominant party in the client-officer relationship when they are dealing with their typical client – the rural



15

female user of the small loan product. This helps keep collection rates high. It is a different matter, however, when the same loan officer deals with the more affluent clients who take the bigger business loans, especially in the urban areas. There, the social conventions may work against the officer: he may be less educated, poorer, and less well dressed than his client, and many SEL clients will expect him to be deferential. Not surprisingly, he finds it much harder to control repayments. Not surprisingly, portfolio quality is poorer for these loans. This is an important but little noticed reasons for the slow growth shown by NGO-MFIs in the middle-income business loan sector. We return briefly to the theme in the final section. In one customer-service issue ASA did take an early lead. It was the first of the big Bangladesh MFIs to recognise that group-based joint-liability (the cross-guaranteeing of each others’ loans by group members) is not a dependable mechanism for ensuring good repayment. It may help collect loans from some weak borrowers, at least in the early months of a group’s life, but at the unacceptable cost of offending good borrowers, who resent being denied loans because someone else failed to pay. This lesson was learned, slowly, through the 1990s, but ASA was the first to react to it publicly and practically: by 1995 it was announcing that ASA no longer used it, and was instructing its loan officers to regard the weekly meeting more as a repayment collection point than as a joint-liability mechanism. This has given ASA a somewhat distorted reputation, especially among other regional MFIs thinking to emulate Bangladesh practice, of being the pioneer of ‘individual lending’ in sharp distinction to Grameen Bank’s ‘group lending’. This misinterprets what is actually found on the ground in Bangladesh, where this distinction between ASA and Grameen is barely perceived by microcredit clients, since the pattern of the weekly meetings of the two organisations, and of that matter of almost all MFIs, is very similar. Moreover, Grameen Bank now not only insists in its publicity that it does not use joint liability, it even asserts that it never did so. 5 Operational management ASA’s sharp focus on the basics of its business – disbursing and recovering small loans – is reflected in every aspect of its operational management practices. Reporting systems: A majority of the most senior staff at head office, including three of the four executive vice presidents, and most of the directors (including deputy, assistant, and junior assistant directors), travel frequently to the field, spending half or more of their time there. This is because head office, besides housing divisions carrying out routine tasks (accounts, audit and taxation, finance and MIS, and so on) operates largely as a machine for gathering observations from the field and processing them into decisions that can be quickly operationalised. Back in their offices, besides writing up their own visit reports, the directors spend another large part of their time analysing reports submitted by field staff, of which the most important by far is the District Manager’s Monthly Report. A glance at this report format reveals the issues that head office takes most seriously, and there are few surprises: they focus on performance as measured by portfolio size and quality. The two-page format begins with the figures for changes in branch and staff numbers, and plan-versus-outcome numbers for clients, disbursement and recoveries. This is immediately followed by a narrative box on ‘Overdues Management’ in which district managers record the steps they have taken in this respect since the previous report. Then come two tables: in the first DMs are required to list the district’s ‘Most Troubled Branches’ (in terms of overdues or poor outreach) and to state what they have done about them, and in the second to record findings from the nine branches they visited in the month in the course of making their contribution to the ‘cascade’ of branch monitoring visits we described in an earlier section. Once again, ASA is systematically and relentlessly bearing down on the essentials – getting loans disbursed and then getting them collected. Procedures and compliance: Further description of head office behaviour will show how the Manual, which controls procedures and provides the yardstick for compliance, is created. Through casual and planned meetings, gossip, and the writing and reading of reports, issues that appear to matter emerge from the mass of observations, and reach the EVPs and the President, who will decide on the best line of action for some, and leave others for further discussion. At the Monthly Policy Making Meeting, they are all reviewed in an open forum. We observed a meeting in early November 2007. There were 50 head office staff in the room (all male except for the President’s secretary, who took minutes). The agenda had



16

28 items on it, not arranged in any particular order, including some carried over from the previous meeting. They ranged from the smallest detail (if field staff leave within six months of being issued with a raincoat, their final pay will be reduced to recover part of the cost of the coat) to major decisions (to suspend the opening of new branches for six months). Choudhury chaired. Although almost everyone spoke at some time, the meeting took only 90 minutes and all 28 items (along with an additional five that were raised at the meeting itself) were dealt with, either by a decision or by being referred to ad hoc committees that were charged with bringing a recommended solution to the next meeting. Choudhury closed each item by indicating he was satisfied with the discussion and summarising the conclusion. Within days a circular was prepared, recording and turning into instructions the decisions that affect the field, and sent to every office: branch managers have a stack of these on their desks. At two yearly intervals the Manual is updated in accordance with the circulars. The system is unsophisticated, but simple and effective. Every senior officer is involved, at least as an observer, in every decision, from the grand to the trivial: no-one can later claim that they had no opportunity to challenge it. That silences in-house criticism to a useful extent. Everyone listens in on details that may not affect them directly, and this helps to prevent the emergence of independent silos of knowledge and responsibility. Just by being at the meeting, all senior staff acknowledge, implicitly, the leadership of the President. That undoubtedly helps to preserve cohesion and to ward of challenges to the leadership. The meeting allows everyone to display their skills of persuasion, and have it measured by the weight of decisions. That helps prevent particular issues turning into battlegrounds. Finally, each long agenda is a reminder to everyone of what really matters to ASA, as echoed in the issues that come up most often – cost control, growth and avoiding overdues. Internal control: Because ASA’s extreme decentralisation means that almost every transaction is carried out by branch staff, internal control measures can be focused there, and not much diluted by having to inspect books and behaviour at offices further up the hierarchy. As we have seen, the ‘cascade’ of visits that a branch receives from superior officers is the principal ongoing tool of control. But it is supplemented by other measures. Loan officers swap clients at regular intervals as short as six months, leaving little time for a rogue officer to form the sort of cosy relationships with clients that can lead to ghost loans and fictitious savings withdrawals. Misappropriation, when discovered, is punished quickly, severely, and peremptorily – Bangladesh does not have employment codes that make it hard for organisations to fire staff. Simple products with predictable cash flows provide much public transparency: famously, and as illustrated in the box on an earlier page, each branch has a blackboard on which, each morning, the expected total ‘take’ for each loan officer is inscribed, and a loan officer who comes back with some other amount needs to explain it in detail to the manager. The new computer programme will emulate this system, presenting a review of the day’s expected incoming cash flow to the assistant management each morning when the machine is switched on, as well as reminding him what loans need to be issued that day. Finally, where things are bad – and ASA has some whole districts as well as smaller geographical areas where performance is poor – senior staff are allocated sets of branches that they must visit more often than others, complete more detailed reports, and talk to more clients. Credit management: New clients receive a first loan of a set value. Clients in good standing (with up-to-date loan repayments) are automatically eligible for a follow-up loan as soon as they repay the previous one. The value of the first loan, and of the subsequent increment, is set by directors, for each district or in difficult areas for each branch. Unless the loan officer has spotted something unusual (preparations by the household to leave the area, for example), no further underwriting tasks are carried out. The loan term and the repayment schedule are fixed and non-negotiable. A notional ‘loan use' is recorded but rarely taken very seriously: staff know that these loans, which are small relative to household income (rarely more than a quarter’s income and usually much smaller) enter the household cash flow and get repaid from it. The repayment record is a much better indicator of ability to pay than any other, which is why clients who do fall behind are dealt with quickly – usually by being chivvied to repay their current loan and then being eased out of membership of the client group. We have seen that, to an increasing extent, loans are backed in part by savings and security fund deposits, so in many cases a troubled loan can be cancelled, and the client removed from the list, by simply debiting the loan from the deposits. Each



17

month in ASA, between 120,000 and 150,000 clients exit, with somewhat more (about 140,000 to 170,000 at present) joining14

Working hours for field and for head office staff are very different. In the field, as we have seen, staff regularly work a very long six-day week, busy from early morning (when they go out to group meetings)

. Staff work hard to prevent the tell-tale overdues from occurring in the first place. Loan officers may use their superior status to pressure clients to pay at the weekly meeting, sometimes using techniques that come close to social collateral – for example by keeping everyone sitting in the meeting until someone lends the overdue client enough money to make at least a token repayment. If that fails, the loan officers make home visits, and if they still have no success, two or more of them will go to the field again, after dark, and camp on the doorstep of an overdue client until they get the cash. This is the main reason for the very long hours that branch staff work – often up to 10 or 11 at night. ASA is not alone in this: it is standard procedure for MFIs in Bangladesh. These ‘bicycle calls’ are perhaps more effective than the call-centres and payment-reminder calls made in wealthier environments, because they are done in public, and for Bangladeshis it is particularly shaming to show inattention to guests on the doorstep. For a small number of loans – the SEL or ‘small entrepreneur lending’ loans that account for about 10% of portfolio and less than 4% of clients – there are appraisal formats that record business inventory, cash flows, and credit lines, and ASA takes pre-signed advance-dated repayment cheques drawn on other banks. By and large, ASA is not yet as good as doing this kind of business as it is at the small loans sector, as poorer portfolio performance numbers show. Human resources: Like many Bangladeshi NGOs that were formed in the wake of the independence war of 1971, ASA is still led by its founder, and many of the most senior positions are held by people whose history with the organisation goes back to the its early days. If we take the four executive vice chairman, for example, we find that the last to join ASA did so in 1993, just after ASA went into microcredit. The next most senior layer is dominated by old hands. Such histories lend organisations a particular flavour: a sense that the leadership is immutable, and a tendency to be wary of well-educated newcomers, for example. This has been challenged recently by the decision to computerise, which has brought a new sort of employee into ASA: young people with high level technical qualifications, with good formal logical skills, better at English, and able to negotiate for high salaries. There has been some awkwardness in absorbing them: for example, their contract conditions have been changed as ASA tries to decide whether it wants to have these newcomers as fully regular or as contracted staff. In the head office, there is a palpable gap between them and older staff. ASA has about 24,000 staff members, more than half of them loan officers and most of the rest field-based managers who started as loan officers. Recruitment at ASA was famously simple: up to 50 applicants were given a group interview and the successful ones were identified on the spot. They were then sent for 10 days unpaid experience gathering at a branch, and if the branch manager made no objection, then started paid work as a loan officer at another branch. Recently, as we have seen, this has changed: the recruitment has been decentralised to the district or even the regional level, to speed up recruitment and make it even cheaper to organise – though recruits do now get paid for their probation period. Promotions, too, are handled locally, and loan officers make their way up through the hierarchy mainly by ensuring they do nothing seriously wrong: expansion has been fast enough to permit fairly swift advance. Recently, for example, 3,300 loan officers were promoted to the new position of assistant branch manager. Though there are formal performance assessment procedures (formats that supervisors complete) selection for promotion is usually an informal matter, based on the opinion of local seniors and reflecting the basic skill required – loan repayment collection. Staff at all levels rarely receive formal training. An exception is running now, as assistant branch mangers are trained to use the computers. At head office, there is no systematic in-service training for staff.

14 More and more, joiners are ex-clients. Such clients may drop out and rejoin for all sorts of reason, but it is not uncommon for a client to climb the loan-value ladder, get into trouble, drop out, and then start again with a small loan. These days, even those who are not ex-ASA clients, are current or past clients of other NGO-MFIs.



18

to late in the evening (when they come home from their ‘bicycle calls’ on late payers). In head office the working hours are 8.30 to 4.30 for 5 days a week, and 8.30 to 12.30 on a sixth day – and most people go home on or soon after time. Despite this, ASA finds it not too difficult to send several senior staff abroad at a time, sometimes for many months, to look after its overseas contracts to train local MFIs in ASA’s systems, or to work with new partners or new replications of ASA15

15 Under ‘Catalyst Microfinance Investors’: see

. Shafiqual Haque Choudhury is 61 this year. Fazle Abed at BRAC and Muhammad Yunus at Grameen are a little older. None of their organisations, as far as the public can see, has a clear succession strategy in place, though Abed at BRAC has stepped down as CEO and moved to the chair of the board. Succession, therefore, is a challenge which has not disturbed ASA up to now, but will have to be faced in due course.

www.catalyst-microfinance.com for information

http://www.catalyst-microfinance.com/�



19

PAST, PRESENT AND FUTURE: AN OVERVIEW

theme of this paper has been that ASA’s period of rapid growth – from 1992 to 2007 – was free of many problems that other fast-growing MFIs face for two main reasons. First, ASA chose to enter a

vast market that was only just being opened up, and did so with a single simple product that required little modification for the sixteen years. Second, many of the serious problems of developing a microfinance programme had either already been dealt with before the sixteen years began – or remain still to be faced. Thus, in the 1980s ASA settled its governance and leadership, moving from a utopian ‘people’s organisation’ to a regular NGO format with a compliant governing body, and seeing off challenges to its founder and leader, Choudhury. It built up good relationships with and large flows of grants from donors. It established offices in many parts of the country, and learned how to employ staff to organise poor rural households into groups. It was able to watch microcredit under development by others – crucially Grameen and BRAC – and dabble itself with loans. It worked through its own doubts about the meaning of development and decided for once and for all that credit is what poor people wanted and what poor people thought would most help them to help themselves. When all this crystallised – above all in Choudhury’s mind – ASA moved swiftly. It told its mostly male group members to send their wives or daughters in their place next week, to receive credit, and told its mostly female local fieldworkers to give way to newly recruited young men. It set itself a target – self-reliant growth – which it first met and then went far beyond without needing to change the iconic chart that symbolises it (page 13). After sixteen years of growth, ASA is undoubtedly in a strong position. Its clients have grown to more than 6 million, its loan portfolio to the equivalent of about $355 at current market exchange rates (end 2007), and its adjusted return on assets to 14.5% (2005). In 2001 ASA stopped taking grants from donors, and a little later, in 2003 stopped borrowing from the national microcredit fund wholesaler. It is cash-rich, with mounting retained earnings. In 2005-6 it built its 14-floor commercial tower in Dhaka, financed entirely from its cash balances, without taking any loans16

But microfinance in Bangladesh may be undergoing changes that will require ASA to face up to new problems. It is not easy to say exactly when that will happen, for it has proved too easy to be gloomy about Bangladesh microcredit. Around the turn of the century research reports warned that increasing competition between MFIs, growing numbers of clients with loan accounts in several MFIs at a time, and declining repayment performance mean that the market for the standard microloan is maturing, and that difficult choices lie ahead

.

17

Still, after opening 326 new branches in 2005, 640 in 2006, and another 400 in the first three quarters of 2007, ASA suspended the opening of new branches for at least six months from November 2007. The suspension was caused in part by the need to recruit new loan officers to replace the 3,300 promoted to assistant branch managers to run the newly-installed branch-level computers. But there is a darker side. More than ever before ASA’s leadership is worried about the declining performance of its core product, the annual small loan. The circular that went from ASA to its field offices in mid 2007 did more than tell them not to open new branches. It included a section called ‘measures against further overdues’. Among these were reducing the client-worker ratio, making the opening of a long-term savings account mandatory for borrowers of loans worth 15,000 taka ($200) or more, and new guidelines on how staff

- but then ASA and other MFIs went on to experience several years of their fastest ever rates of growth. It wasn’t so much that new hitherto-unserved locations were discovered and colonised – though some of that did happen – but more that each provider wanted to make sure that it had a presence in every location where its competitors were active. This of course had the effect of increasing local competition and encouraging yet more multiple account holding by clients

16 It has found good quality commercial tenants for most floors, and occupies the two topmost for itself. In 2006 ASA University, a private school, financed by ASA, opened on another three floors. 17 See for example Graham A.N. Wright, Robert Peck Christen and Imran Matin, ASA’s Culture, Competition and Choice: Introducing Savings Services into a Microcredit Institution: MicroSave 2001; and a little later Fernando and Meyer, in the ADB newsletter, op cit, 2002

A



20

should follow up arrears. The circular notes that these new measures must be pursued even if they result in some clients closing their accounts. A similar circular was sent to the field by another large MFI whose directors we interviewed. What choices will ASA have to make if the growth of demand for small annual loans does indeed grind to a halt, with competition between MFIs suddenly intensified as their expansion-spree suddenly comes up against the buffers? First, ASA will seek to protect its share of the current microcredit market, and it is prepared to spend money to do that. As the recent circular shows, it may cut down on the numbers of clients per loan officer in order to give the latter more time to deal with overdues, and it is willing to shed some clients in order to keep good ones. At head office, discussion is already underway about pricing policy: both reducing the price of loans again (ASA already did so in early 2007), and raising rates on savings. ASA has noticed that it is getting harder to attract good new recruits, so improvements to the pay and conditions of loan officers is another possible move. Then, ASA will think about how it might improve the current product. It could adopt some of the advances that Grameen has made since it introduced ‘Grameen II’: variable loan term lengths, and variable repayment schedules, and the right to ‘top-up’ loans half way through the repayment schedule, for example. The current computerisation may help to make these possible without moving too far from the ASA ‘predictable cash flow’ principles. On the deposits side ASA will find it harder to compete with Grameen because Grameen’s formal license entitles it to offer savings accounts for households outside those represented by borrowing-group members. Hitherto this has not been an issue for ASA: it’s earnings from loans are so strong that it has not needed to seek additional deposits18

18 Though these are substantial – see Chart 6, page 7

to fund its lending, and it has not had any other reason to want to work with non-members. But both these factors could change, pushing ASA to seek a legal identity that would allow it to take savings. Moreover, Choudhury has expressed concern that the growing pile of retained earnings at ASA is, under Society Act legislation, effectively owned by no-one. He is known to prefer a more broadly based and more commercial ownership structure. Having a major microcredit NGO pushing for new microbanking legislation would be a big change in Bangladesh, likely to speed up the re-writing of the country’s supervisory system. The other obvious direction for ASA is towards loan products for genuine micro-entrepreneurs. As we have seen, it has tinkered with this over the years, without ever finding exactly the right way to do it. Its latest version – the SEL (small entrepreneurship loan) – hints at what it understands would be necessary: individual loans using no groups, and some basic underwriting appraisals of the business. But so far the performance of these loans has not been quite as good as for the standard small loan, and to overcome that ASA may have to employ a different kind of loan officer and spend a lot more time and money on training them. That would be a big challenge. These challenges – transforming to a new legal identity, making the basic loan more flexible, developing a range of deposit services for the general public, and figuring out how to give and collect loans to individual businesses – are all bigger than anything ASA has had to face since it first shifted to microcredit in the early 1990s. Meanwhile, two other direction are opening up in ASA, though neither of them is an alternative business plan if the volume of small-loan work does decline sharply. The first is non-financial services. The opening of ASA University in 2007 – a for-profit entity but through which ASA also subsidises, via scholarships, the education of children of its microcredit group members – may prove to be just the first of a number of ASA-financed services. A medical college, a string of hospitals at district level, and high schools (and maybe later primary schools) could follow: all are being discussed and in at least one case plans have been submitted for government approval. Sometimes these institutions are talked of as a way of ‘giving back to our members’ the surpluses that ASA has created through microcredit.



21

The other growth area is work overseas. ASA has for many years been involved with developing microcredit outside Bangladesh. There have been commissions from international development organisations to advise Asian and African providers, and other kinds of partnerships. More recently, ASA has been a partner in the creation of ‘Catalyst Microfinance Investors’ (MCI, see www.catalyst-microfinance.com), a fund whose business plan has been evolving, but which is now looking seriously at financing – directly or through local partners – a series of MFIs in countries with big populations of poor people, such as India, China, Vietnam and Nigeria. The assumption is that the basic ASA approach – centred firmly on the no-frills small loan, repaid in frequent instalments – will find a big market in such countries. Choudhury argues that the difference between this fully commercial venture, and the commercialisation of microcredit in Latin America and eastern Europe, is that the ASA replications will demonstrate that lending to the poor, as opposed to microentrepreneurs, can also form the basis of a sound commercial microcredit business plan. One side-effect of this development is that job opportunities will open up for many ASA employees overseas, helping to forestall any blocking of the career path that might have occurred as a result of slower growth at home. Many staff even down to branch manager level are now busy practising their English language skills in the hope of being chosen for such a posting. We may now have reached the point at which ASA in its home market needs to go beyond the basic microcredit product, even as, overseas, it introduces it to tens of millions of hitherto unserved poor households.





22

APPENDIX: REVISITING THE PANKAJ JAIN PAPER

n 1997 the Indian researcher Pankaj Jain wrote Managing Fast Expansion of Micro-Credit Programs: the Lessons from ASA (Dhaka, ASA, 32 pages). We have been asked by MicroSave to comment on its

continuing relevance. Readers of the present paper will find many similarities with Jain’s analysis of ten years ago. In writing about the ‘core lessons’ that can be learned from ASA, he divides them into two ‘domains’: lessons that dispel myths about what is needed to run successful microcredit programmes; and lessons that reveal critical factors that enable fast scaling-up. Dispelling myths: Jain identifies two myths which the ASA case helps to dispel:

1. Group guarantees as social collateral He noted, as we have done, that ‘joint liability’ for loan repayments (or ‘group guarantees; through ‘social collateral’) is not the key to ensuring good repayment performance.

2. The need for participation by beneficiaries in the design of programmes Jain argues that the ‘marginal to non-existent’ role that ‘people’s participation’ plays in ASA’s programme decision-making goes against the ‘axiomatically accepted’ view that ‘empowered local groups have to be an integral part of any development programme’. Jain sees ASA’s group members very much as ASA does, and as described in the present paper: they are customers of a service, and influence its development by choosing whether and how to use it.

Fast scaling-up: Turning to the ‘elements of effective and fast scaling up’, Jain groups them into four categories: decision making; organisational and staff learning; resource acquisition; and organisational development. Each category has a number of key lessons. Under Decision making Jain identifies many of the aspects of the ASA that we have noted in this paper and which are as true of ASA today as they were when Jain was writing. These include: the light ‘decision load’ on field officers, who are given simple standardised instructions to work

to: Jain stresses that this greatly minimizes errors in decision-making simplified administrative and accounting procedures: Jain notes the standardised instructions that

enable ASA to leave staff to get on with their work without requiring repeated approvals unambiguous indicator-based staff performance review: Jain notes, as we do, that these are based

on the specific targets given to branch offices (we have commented that outreach, disbursement and repayment are the main indicators used)

quick, close-to-the-field problem solving: Jain noted, as we do, that ASA’s high levels of decentralisation and standardisation, combined with the pressure of clear performance indicators, encourage staff to tackle problems quickly, before thy become large

responsive rather than participative decision making: commenting on the frequency with which head office staff visit the field, Jain notes that ASA’s decision-making is not directly participative but is able to respond quickly to field realities: this allows decision-making to be quick and relevant: this conforms to the situation we describe in section 5, operational management

Four further lessons are noted under Organisational and staff learning. Again, these remain both true and relevant. no complex skills required for ‘information-carrying’ and analysis: Jain here refers to the effect

that standardisation and simplicity have on the intellectual load placed on staff. ASA does not have to rely on its staff having high levels of reporting nor of analytical skills

similarly, no complex skills required for social organisation: Jain’s point here is that since ASA’s groups function as little more collection points staff do not need highly developed skills in group management: Jain argues that this factor helps ASA keep staff training quick and low-cost

decision making is done at the top level: this reiterates the last point in the previous group, and Jain stresses that confining decision making to the most senior staff relieves ASA of needing to employ and train junior staff with decision-making skills

I



23

senior officers’ exposure to the field: this too repeats an earlier point, and is made again here to stress that requiring seniors to visit the field regularly frees ASA from alternative more complex and costly information systems

Under Management of resource acquisition, Jain made two points. He was writing at a time when ASA was still largely dependent on donors but was just beginning to use bank loans. ASA presented itself to the donor community as cost efficient rather than novel, offering donors

the chance to invest in microcredit, which they already knew to be desirable, but also to place resources with an organisation that would guarantee especially efficient use of their resources. This may be largely true, although it is worth noting that ASA held on to several of its traditional donors whom it had to work hard to convince of the shift to microcredit

ASA was willing to go for funds from existing financial institutions. ASA was indeed the first Bangladesh NGO-MFI to take a commercial bank loan, which it did by mortgaging its head office building

Finally, under Management of organisational development, Jain makes a mixed bag of four points. ASA does not have to deal with complex group dynamics: this repeats a point already made start simple, then grow more complex: here, Jain refers mostly to the product offerings. market-competitive staff incentive policy: this refers to what Jain saw as ASA’s greater reliance,

compared to other NGO-MFIs, on offering competitive pay and conditions to staff as incentives for good performance. This point may have diminished in importance since: ASA’s pay and conditions for field staff have tended to move close to the average for NGO-MFIs

insulating field staff from social pressure: most field staff come from outside their home districts, and, being mostly young men, can be housed at the branch office building. Jain believed that this helps to insulate them both from distractions from their own families, and from influence by local interests. This point may remain largely true but is softening somewhat as ASA is becoming more prepared to allow both men and women to work in their home districts

In summary, Jain’s analysis was perceptive and accurate, and much of it remains valid ten years later.

MicroSave...As the chart shows, ASA began opening urban branches in 2001, and the number has grown steadily. One additional function that urban branches perform is to members who migrate

Documents